Drug use and abuse can substantially alter human behavior and seriously impair the wide variety of skills necessary for work, school, and normal everyday activities. Although national research efforts, such as the National Household Surveys of Drug Abuse (NHSDA), provide valuable information on the prevalence of drug use in the general public, few studies have thoroughly examined the relationship between drug and alcohol use and labor market outcomes. In this study, we propose to use data from the 1985, 1988, 1990, 1991 concatenated public-use file of the NHSDAs to examine worker prevalence of drug and alcohol use in the workplace and to assess the multivariate relationship between drug and alcohol use and labor outcomes. Specifically, we will complete the following tasks: 1. Produce prevalence estimates for workers for each year of lie concatenated public-use file, by age, gender, race, income, and other sociodemographic groups. 2. Analyze patterns in worker prevalence estimates and describe and compare changes in prevalence overtime. 3. For each NHSDA, estimate the cross-sectional relationship between drug and alcohol use and five labor market outcomes: earnings, days absent from work, number of weeks worked, whether the respondent worked in the last year, and whether the respondent is currently employed, controlling for sociodemographic characteristics. 4. Estimate how these cross-sectional relationships between drug and alcohol use change between 1985 and 1988, 1988 and 1990, and 1990 and 1991. The paucity of scientific information on the consequences of drug use in the workplace hinders decision makers in assessing the magnitude of the costs associated with substance abuse and in establishing appropriate treatment programs. The proposed study will establish empirical regularities that will greatly expand our knowledge of the relationship between drug and alcohol use and labor market outcomes. The findings will help policy makers determine how and where to allocate their scarce resources.